Wearable devices like smartwatches and fitness trackers have revolutionized the way we monitor our health. Worn around the clock, these devices quietly collect valuable data—from heart rate and blood oxygen levels to sleep quality—giving users a real-time window into their well-being without disrupting their daily lives.
A recent study has found a way to track heart rate without requiring people to wear anything at all. Instead, all they need is for someone to use their smartphone, which is quite convenient, since the average person already spends upwards of five hours a day on their phone.
This technology, supported by deep learning, captures an eight-second video using the front camera every time a user unlocks their phone. The video information is then used to passively monitor heart rate from a distance by detecting subtle changes in skin color caused by blood flow. To ensure participants’ privacy was not compromised, videos remained on their devices until they were manually reviewed and approved for upload.
According to findings published in Nature, the system not only met industry accuracy standards, keeping heart-rate measurement errors below 10%, but also maintained its accuracy across all skin tones—a win, given that many existing technologies perform poorly on darker skin tones.
Keeping a close watch on the heart
The heart does the crucial job of circulating oxygen and nutrient-rich blood throughout the body to keep all the cells able to respire and perform their specific tasks. Heart rate, especially the resting heart rate , provides a window into how the heart is functioning, as well as our overall health and energy levels. Measuring resting heart rate in a clinical setting requires a person to rest for prolonged periods, which makes long-term tracking difficult.
Wearables solve this by passively collecting heart rate data throughout the day, when one is resting or sleeping, to estimate RHR, thereby revealing useful changes in cardiovascular health and illness. However, such wearables are not accessible to most people; fortunately, smartphones are.
Studies show that 69% of adults globally and 90% in the US own a smartphone, with an average of 144 interactions with the device each day.
For this study, the researchers recruited 696 participants with diverse ages, sexes, and skin tones. More than 192,000 videos from 485 participants were used to train the neural network for their new passive heart-rate monitoring (PHRM) system. It measured the heart rate using a technique called remote photoplethysmography (rPPG).
With every heartbeat, a surge of blood passes through the vessels in the face, slightly altering the amount of light reflected by the skin. These changes, though invisible to the naked eye, can be captured on a smartphone’s camera, allowing the system to measure a person’s heart rate.
After the system was up and running, a separate group of 211 participants was used to validate its performance in both laboratory and real-world environments.
The data indicated that a smartphone can accurately measure heart rate under different lighting conditions. The team tested PHRM against reference electrocardiograms and found heart rate measurements fell within 10% error across all skin tones and met industry accuracy standards.
By stitching together several short measurements across the day, the system estimated daily resting heart rate nearly as accurately as a professional-grade wearable strap.
Once privacy concerns are addressed and the technology is validated in larger populations, a system like PHRM has the potential to put heart-health monitoring in the hands of billions, including communities lacking access to traditional health care systems. The ability of smartphones to measure heart rate also leaves us with the question of what other health signals could we monitor using our phones?
Written for you by our author Sanjukta Mondal, ed